Post on 14-Jul-2020
transcript
1
TangibleCircuits: An Interactive 3D Printed Circuit Education Tool for People with Visual Impairments
Josh Urban Davis1, Te-Yen Wu1, Bo Shi1,3, Hanyi Lui1,4,
Athina Panotopoulou1,2, Emily Whiting2, Xing-Dong Yang1
Dartmouth College1, Boston University2, Beijing University3, Tsinghua University4
{josh.u.davis.gr, te-yen.wu.gr, athina.panotopoulou, xing-dong.yang}@dartmouth.edu,
bs199857@163.com, lu-hy15@mailes.tsinghua.edu.cn, whiting@bu.edu
ABSTRACT
We present a novel haptic and audio feedback device that
allows blind and visually impaired (BVI) users to understand
circuit diagrams. TangibleCircuits allows users to interact
with a 3D printed tangible model of a circuit which provides
audio tutorial directions while being touched. Our system
comprises an automated parsing algorithm which extracts
3D printable models as well as an audio interfaces from a
Fritzing diagram. To better understand the requirements of
designing technology to assist BVI users in learning
hardware computing, we conducted a series of formative
inquiries into the accessibility limitations of current circuit
tutorial technologies. In addition, we derived insights and
design considerations gleaned from conducting a formal
comparative user study to understand the effectiveness of
TangibleCircuits as a tutorial system. We found that BVI
users were better able to understand the geometric, spatial
and structural circuit information using TangibleCircuits, as
well as enjoyed learning with our tool.
Author Keywords
Tangible User Interfaces, Universal Design, Accessibility,
Circuit Prototyping, Education Tools
CSS Concepts
• Human-centered computing~Human computer
interaction (HCI); Accessibility Systems and Tools,
Haptic devices; User studies;
INTRODUCTION In the maker community, novices learn circuits with
breadboards by following examples in tutorials from the
web. However, most of the existing web tutorials are
inaccessible to the blind or visually impaired (BVI)
community because they rely heavily on visual information
to communicate the material (Figure 1B). This is a
significant loss considering that members of the BVI
community have traditionally been inventors of life-
changing electronic devices that benefit both blind (e.g.,
Optacon) and sighted people (e.g., cruise control) [18, 20].
The high bar of entry to learning electronics excludes the
BVI community from participating in innovation via making.
BVI children also miss-out on critical STEM education and
further high-tech careers [10.]. While many accessibility
tools exist, most do not encourage or enable BVI users to
create their own accessibility tools.
E1: “Blind people are born makers because the world was
not made for them. They have to recreate the world for
themselves to thrive.”
Thus, it is our vision that these tools must be designed to
support learning for BVI users, enabling and unleashing their
creative potential. In this paper, we propose an interactive 3D
printed tutorial system, TangibleCircuits, that combines a
cost-effective tactile model of a breadboard circuit with
audio-feedback for BVI makers and students.
TangibleCircuits comprise an automatic parsing tool which
Figure 1: Overview of the TangibleCircuits system. A) 3D model parsed from Fritzing Diagram; B) Fritzing Diagram used as input by the
system; C) User interacting with the 3D printed model’s audio and tangible feedback.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. Copyrights for
components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to
post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from Permissions@acm.org.
CHI '20, April 25–30, 2020, Honolulu, HI, USA
© 2020 Association for Computing Machinery.
ACM ISBN 978-1-4503-6708-0/20/04…$15.00 https://doi.org/10.1145/3313831.3376513
2
translates a circuit diagram (Fritzing format) into a 3D model
that is printable with a commercial 3D printer and Proto-
pasta Composite Conductive PLA material. The tactile
circuit model has components printed using conductive
filament and can be affixed to a smartphone to allow for
touch-based interaction for learning. When each component
or wire is touched, audio feedback details the name of the
component, the position, and other details regarding its
connection and implementation.
TangibleCircuits is intended to broaden the inclusivity and
accessibility of maker spaces and engineering classrooms by
allowing instructors to create cheap, portable, and easy to use
multimodal circuit tutorials. Our vision for TangibleCircuits
is to allow tutorial authors and instructors to generate a
tangible model and audio interface from existing Fritzing
diagrams. These resulting tools can then be 3D printed using
a commodity 3D printer and affixed to touch-screen devices
to serve as multimodal accessibility tools for BVI students.
TangibleCircuits was developed with a user-centered
universal design approach, where a series of studies were
conducted to understand the problem space and its
magnitude. To begin our investigation, we conducted a semi-
structured interview with 3 BVI makers in order to
understand several major accessibility issues they
encountered using electronics education tools. Examples
include the difficulty in understanding the spatial
(component layout), structural (debugging), and geometry
information (i.e., component size and shape) of breadboard
circuits. In addition, we evaluated the magnitude of these
issues by surveying 3910 online tutorials from the most
popular open-source tutorial platforms (Arduino Projects
Hub and Fritzing Hub). Online tutorials were examined due
to their common use as teaching material for novice
engineers and makers. We found that that over 98% of online
tutorials were not adequately accessible to BVI users
according to the Web Content Accessibility Guidelines
(WCAG) [5]. From these preliminary investigations, we
extracted a series of design guidelines for TangibleCircuits.
To evaluate the effectiveness of our approach, we conducted
a user study with 8 self-reported blind and 6 visually
impaired/low vision/legally blind participants, where we
evaluated the accessibility of TangibleCircuits and web
tutorials modified to be BVI accessible according to WCAG.
We found that our system was better at assisting BVI users
at recognizing the geometric information, spatial and
structural information of components within the circuit.
Participants also discussed that TangibleCircuits was fun to
use, and significantly less strenuous and frustrating to
interact with than online web tutorials.
The main contributions of this work are: (1) an
understanding of the accessibility issues in the existing
circuit learning tools for BVI users; (2) an approach to
address the issues using interactive tactile models for circuit
tutorials; and (3) insights from a user study, evaluating the
accessibility of our prototype and web-tutorials modified to
meet standards of WCAG web accessibility.
RELATED WORK
This work builds upon many intersecting bodies of work
including Circuit Prototyping and Educational Tools,
Tangible Interactions for Visually Impaired Persons, and
Insights from existing STEM education tools.
Circuit Prototyping and Educational Tools
Prior work has shown that novice users face substantial
difficulty in designing and building physical computing
systems [8, 29]. Some challenges include choosing correct
components (geometric information), wiring components
together (spatial information), and debugging (structural
information). Several research systems have been developed
to address these challenges. For example, Toastboard [11] is
an intelligent breadboard that assists novices with debugging
through LED indicators on the board itself, and a software
interface that provides troubleshooting tips. Other systems
teach fundamental concepts of circuit design, and
programming. For example, Programmable Bricks [30]
allows children to develop electronic hardware using LEGO
bricks embedded with computers, sensors, and actuators.
Finally, a number of systems have been developed that aid in
sensing the state of the electronics components in embedded
systems [6, 13, 27, 40, 42], data which could aid in
debugging and troubleshooting.
Unlike the systems focused on developing novel hardware and sensing techniques, our work examines how insights
from these techniques can be adapted to enable visually
impaired persons to learn electronic prototyping. It is our
intention to create a platform that simultaneously employs a
universal design approach, as well as ensures the user can
learn as autonomously as possible. The purpose of a
universal design approach is to similarly enable visually
impaired and traditionally sighted users alike with a single
prototype design in order to encourage the tool’s wide
adoption. Additionally, our goal of ensuring autonomy is to
support the pseudo-autodidactic nature of online learning
platforms. For these purposes, tangible and audio feedback
systems present a viable modality to achieve these goals.
Tangible Interaction for BVI Persons
Most technologies that are accessible to BVI people
substitute visual information with audio-feedback or touch-
feedback. Touch is a promising modality for sensory
substitution, as previous studies have revealed superior
tactile acuity for blind people over sighted people [10].
However, few tangible user interfaces (TUIs) for visually
impaired people have been designed, and the existing
accessible TUIs mainly broaden accessibility to geographic
maps and diagrams. Examples of tangible diagrams include
a prototype for the non-visual exploration of graphs and
maps by McGookin et al. and TIMMs by Manshad et al. [25,
26]. These tangibles systems provide multimodal feedback
for the creation and modification of diagrams and maps.
Other multi-sensory projects include MapSense and
IllumiWear [9, 12] which integrated scents (e.g., olive oil,
honey) or sound, thus creating a multi-sensory map. More
closely related to our interests are tangible maps, where map
3
elements are represented by physical objects which are often
augmented with audio feedback [14]. In some cases, users
can not only explore the maps, but build and modify them by
manipulating and moving the objects. Similarly, the
prototype by Schneider and Strothotte [37] enabled visually
impaired people to construct an itinerary using building
blocks of various lengths with the help of audio cues.
Tangible Reels [15] are physical icons on a multi-touch table
representing points of interests. The system guided the user
with audio instructions to correctly place, link, and retrieve
the names of objects.
TUIs have shown many advantages over standard mouse and
keyboard computer interfaces. They foster collaboration and
have also proved to increase engagement of students in
learning tasks [17, 32]. Moreover, constructing tangible
maps improves the understanding and memorization of
spatial information in the absence of vision [15]. Similar to
our interests is Interactiles which uses conductive 3D
printing to increase smartphone devices accessibility [44].
Some preliminary work has been conducted translating these
tools into the domain of STEM education tools.
STEM Education Tools for Visually Impaired Persons
Designs for learning computer programing and electronic
engineering for BVI users are limited [24, 31, 39]. The few
developments in this area include accessible programing
languages (i.e. Quorum) and speech interfaces (i.e.
Emacspeakiv) that can be effective tools for those who
already know how to code, but are less suitable for novices.
To assist BVI computer science majors to learn how to
program, Smith et al. [38] introduced JavaSpeak, an editor
providing additional information about the structure and
semantics of written Java code. Other examples include
systems which simplify programming logic and provide
audio feedback, and tools which help children using screen
readers create chatbots [7, 31]. Additionally, Kane and
Bigham [21] described BVI teaching students how to analyze
Twitter data, producing 3D printed visualizations that
allowed for a tactile exploration of their program output.
These approaches mostly serve to increase the accessibility
of text-based programing by simplifying coding syntax or
teaching the use of screen-reader or magnification software.
As such, they are more suitable for textual information than
visual information. In addition, with engagements being
primarily bound to a computer screen, they rarely support
hands-on physical engagements. Thus, they do not capitalize
on the possibilities offered by manipulating physical objects
for learning complicated concepts [36], or for supporting
collaborative learning [19].
Despite tangible programing languages and tools for sighted
users [e.g. 16, 19, 28, 3, 41], little work has been done to
explore the effectiveness of these modalities within the realm
of physical computing. Some early work in this field is
evident in the work of Li et al. who used tactile templates
combined with audio feedback to aid users in understanding
and manipulating the spatial information of a web-page
layout [22]. In addition, some early work documents the
potential usefulness of 3D printed models as learning tools
for BVI users [33, 34, 35]. TangibleCircuits builds upon
these insights in order to design a tangible and audio system
for educating novice BVI makers.
STUDY 1: SEMI-STRUCTURED INTERVIEWS
To aid in our understanding of current practices and needs
for accessible circuit prototyping education for BVI
engineers, we conducted a semi-structured interview with 3
BVI participants familiar with circuit prototyping
technologies. This included a blind engineer who facilitates
workshops for BVI people to learn about electrical
engineering, a blind technology administrator at a local
school for the blind, and a BVI student whom had previously
studied physical computing at the college level.
Results
We first wanted to understand current practices in hardware
education and found that web-tutorials were often relied
upon by our interviewed instructors. Literature corroborated
this insight, revealing that web tutorials were commonly
used by educators of a variety of backgrounds as a principle
source of classroom material [13]. This indicated to us a need
to better understand the current accessibility of open-source
tutorial systems (see STUDY 2). One of our experts
expressed frustration at using these tutorial systems within
the classroom. They revealed that upon matriculating into university, their intention was to pursue engineering as a
major, but found that while some accessibility tools made
programming easier, navigating circuit implementation was
impossible due to the cognitive load required to understand
circuit diagrams using a screen reader. This indicated to us
the need for multimodal feedback as a necessary design
consideration (see Multimodal Feedback). We also inquired
about current tangible methods used within this domain, and
found that tactile diagrams were commonly used, but due to
the abstraction used in direct graphic translation, these
diagrams remain largely unusable. Thus, it is imperative for
our design to support recognizability of components more
suitable for a tactile domain (see Support Recognizability).
Finally, we discovered that existing circuit education tools
for producing tactile assets included braille embossers and
swell paper which are prohibitively expensive and not
commonly available in engineering educational settings.
This indicated a need to make our tool as inexpensive and
ubiquitous as possible, and usable to makerspace and
electronic classroom educators with more ubiquitous tools
(see Automate Accessibility).
STUDY 2: EXISTING TUTORIAL ACCESSABILITY As indicated by our interviewees during our semi-structured
interviews, web-tutorials often serve as a primary source for
classroom material for novice engineers. In order to assess
the current accessibility of open-source tutorial systems, we
conducted a formative study of web-based hardware
computing tutorial resources. The focus of this initial study
was to understand the magnitude of accessibility limitations
within online open-source tutorial platforms, as well as
insights into common web accessibility pitfalls.
4
Magnitude of Tutorial Accessibility Limitations
For the purposes of this study, we collected 7321 online
tutorials from two popular online open-source tutorial
platforms: Arduino Project Hub [1] and Fritzing Hub [2].
After filtering, this collection comprises 3109 tutorials
collected from Arduino Project Hub and 801 tutorials taken
from Fritzing Hub. We filtered out tutorials that were empty,
not in English, or were significantly incomplete (i.e. missing
project description). These tutorials were then analyzed
using the Web Content Accessibility Guidelines (WCAG) an
online protocol and guideline system for ensuring web
accessibility [5]. These guidelines are divided into 4 areas of
concentration: perceivable, operable, understandable, and
robust. From these guidelines, we extracted 4 characteristics
of accessibility which are applicable to hardware computing
tutorials. We then assessed the accessibility of the collected
tutorials based on this criteria. Results are detailed Figure 2.
Results
While 79.5% of the 3910 entries from Arduino Hub
contained graphics or photographs, only 2% contained
graphic descriptions. This violates the WCAG guideline of
Perceivable, weakening the accessibility for users with visual
impairments. Furthermore, the preliminary data shows that
53.8% of these tutorials use a circuit diagram and 15.5% use
a schematic, but only 3% contain circuit descriptions (see
Figure 2). An understanding of these visual medias is
imperative to completing the tutorials because the circuit
diagram (i.e. Fritzing Diagram) communicates spatial
(component layout), and geometric (component size and
shape) information of breadboard circuits. Both circuit
diagrams and schematics communicate structural (wiring)
information which is largely missing from these tutorial
systems. Only 13.3% contain written step-by-step
instructions, and less than 1% contain video with captions.
Overall, we found that less than 2% of tutorials surveyed met
the criteria for accessibility according to WCAG, indicating
a significantly limited accessibility in these online tutorial
platforms. Of the tutorials surveyed, we found that the 801
tutorials extracted from Fritzing Hub were less accessible
than those from Arduino Project Hub. Fritzing Hub tutorials
because they relied heavily on circuit diagrams (Fritzing
diagrams) as their primary tutorial material, and lacked
textual descriptions of the circuit or components. In fact,
98.2% of Fritzing Hub tutorials contained a circuit diagram,
but less than 1% contained circuit or component
descriptions. Details from this analysis can be found in
Figure 2. Furthermore, through this process, we identified
key pitfalls of frustration when navigating these media using
screen readers. Component descriptions, for example, if
included in the tutorial, were usually contained within large
HTML tables which were frustrating to navigate using a
screen reader. This was largely due to the tables containing
information pertaining to the operation of the webpage (such
as table indices and tag information) that was not relevant to
tutorial material. In addition, relevant information such as
component names was also inaccessible because component
names were often extracted from the file name of images
associated with the component. This resulted in verbose,
unreadable component names that were difficult to associate
with a given component.
The results of our study suggest that a system designed to
meet this accessibility gap must mitigate the significant
difficulty, time, and labor necessary to communicate
component descriptions and circuit connectivity to a novice
BVI learner. These results motivated us to provide direct
access to component information through 3D replicas (see
Support Recognizability) and audio feedback of a touched
component (see Multimodal Feedback). Furthermore, given
that broadening accessibility to these tutorials is a time-
consuming endeavor, it is necessary to automate as much of
these tasks as possible in order to create a system that is easy
for BVI users to understand the circuit tutorial contents.
DESIGN CONSIDERATIONS
Based upon the insights from the above study, we devised a
series of criteria to inform the design of our system. From
our collected semi-structured interviews and our preliminary
study, we devised the following considerations.
Support Recognizability
According to our initial study, one of the key components
missing from the tutorials examined in our study is adequate
description of components. While most tutorials contained a
list of components, none contained adequate visual or tactile
component descriptions. Furthermore, we learned from our
semi-structured interviews that tactile graphics and maps
were often insufficient due to their direct translation of
abstract graphics to a tangible medium. We thus chose to
explore a direct 3D representation of components for our
prototype. Any system designed to meet these needs must
therefore account for this discrepancy in current tutorial
system technology.
Multimodal Feedback
A recurring limitation in current tutorials lies in the lack of
non-visual communication methods. This discrepancy not
only violates the WCAG Perceivable principle, but also
excludes populations unable to interpret visual material.
Figure 2: Overall results of magnitude accessibility assessment for
3910 web tutorials. Values indicate number of tutorials containing
specified media.
5
Thus, a system designed to account for this limitation must
incorporate multiple forms of feedback and guidance, (e.g.
audio, tangible, etc.) in order to increase accessibility.
Support Understanding of Circuit Structure
A key to understanding the functionality and implementation
of a circuit is understanding the structural and spatial
information of the circuit, including connectivity of different
components and their interactions [8]. According to our
experts, this principle is not present in current hardware
computing accessibility technologies. E2: “Descriptions of
circuit diagrams only get you so far, you really need to see
how things are put together to get them to work…otherwise,
debugging is near impossible”. Therefore, our system must
account for this knowledge gap, enabling users to understand
the layout and interaction of various components.
Automated Accessibility
As evidenced by our semi-structured interviews and
formative study, considerable time and effort is demanded of
tutorial designers to meet standards of accessibility.
Therefore, it is necessary to automate a portion of the
accessibility limitations evident in these tutorials. While the
Fritzing platform enables a wider audience of novice
engineers and makers to create and interpret circuit
diagrams, our previous study indicates that the current
interoperability of this visual media is exclusionary to BVI engineers. Furthermore, current technologies such as tactile
diagrams and maps require accessibility equipment which
may not be common in classrooms. For this reason, we chose
to focus on a system design which incorporated 3D printing,
a more common tool in most electronic educational and
maker spaces. Thus, integrating additional features which
automate the rectification of accessibility limitations in this
platform would broaden the benefits of this enabling
technology to a wider audience.
TANGIBLE CIRCUITS
To account for the above design considerations, we created
TangibleCircuits: an audio and tangible circuit tutorial
system. TangibleCircuits comprises an automated parsing
system which translates a Fritzing diagram from a visual
medium to a 3D model and voice annotation. This model can
be 3D printed using Proto-pasta CDP12805 Composite
Conductive PLA material and affixed to a commodity touch-
screen smartphone or tablet for voice output. The resulting
interactive tactile diagram allows a user to tangibly
understand a circuit using touch-triggered voice-feedback.
Interaction Design Overview
Since audio and tangible feedback have demonstrated
effectiveness for communicating information to BVI users,
our design focuses on integrating these two modalities of
communication. To interact with the tactile diagram, a user
simply touches any component, triggering audio information
regarding that component to be read to the user. This allows
a user to gain insight into the implementation and
composition of the circuit while becoming familiar with the
tangible shape of each component.
Implementation
TangibleCircuits takes a Fritzing Diagram as its input, and
parses the diagram into a 3D model and touch based audio
interface. These two complementary components comprise
our tutorial system, and the resulting interactive tactile
diagram operates on a commodity capacitive touch-screen
device, such as a smartphone or tablet without any
modification to the device.
Audio Interface
The audio interface consists of a series of buttons laid-out on
the display of the touch-screen device (See Figure 4). Each
button is associated with a different component present
within the circuit diagram. When touched, the device reads
audio information related to the target component associated
with the button. This information includes the target
component name, relevant neighboring components to which
the target component is connected, and implementation
instructions such as when the component should be inserted.
The system repeats this information until the user releases
the button, and only responds to a single touch. The user is
notified if they are touching more than one component.
3D Printed Circuit Model
The 3D model is extracted from the Fritzing Diagram and
renders an approximate replica of the components within the
circuit. This is intended to provide a tangible approximation
that mimics the tactile qualities of the physical breadboard
circuit the model represents. TangibleCircuit’s components
and wires are printed using Proto-pasta CDP12805
Composite Conductive PLA hard extrusion filament. This
conductive filament is crucial to the operation of the device.
The breadboard and case is printed separately with non-
Figure 4: Audio interface of TangibleCircuits displayed on a
commodity smartphone.
Figure 3: A user interacting with the TangibleCircuits prototype
6
conductive PLA filament. For our purposes, and for the
purposes of reusability we printed the case and circuit
separately. These two elements can easily be printed as a
single unit as needed. Although the case and board are
printed using non-conductive filament and the components
printed with conductive filament, both these elements were
printed as a single unit using a multi-material 3D printer.
This casing allows the tactile circuit diagram to sit above the
capacitive touch screen. Each component in the tactile circuit
diagram sits directly above its corresponding audio interface
buttons, and thus triggers the voice annotation below each
component when touched (See Figure 5). The resulting
interaction allows for both audio and tangible interaction to
inform the user of the circuit’s spatial, structural, and
geometric information.
Automatic Parsing
In order to reduce the labor required to create the necessary
audio interface and 3D model for each circuit, our system
includes an automatic parsing tool which renders the related
3D model and audio interface for each circuit. Our parsing
tool first takes a standard Fritzing Diagram as input , which
is then parsed for component id tags. These id tags contains
the name of the component, its x and y coordinate position
within the diagram layout, the pins of the breadboard in
which the component is inserted, and the ids of the
components connected to the target component. Wires are
also described in a similar way, in that their id contains their
pin insertion locations and connected components. Since
each component and wire id tag contains a series of x, y
coordinate positions as well as pin insertion locations, we are
able to determine the relative size of the component as well
as its relative position on the breadboard. Our system then
identifies a 3D modeled component within our component
dictionary, comprising a series of component ids and their
corresponding 3D models. These component models were
taken from open-source online repositories and collected into
our dictionary. Once this has been completed, our tool
assigns each 3D model component to a location on a 3D
breadboard according to the x and y coordinate positions
associated with the parsed component id. The 3D model is
then rendered and output as an stl file for 3D printing. The
corresponding audio interface is parsed in a similar manner,
where each component is assigned to a touch button whose
size and location are determined by the two x and y
coordinate positions of each component id. In addition, each
component id tag contains information regarding the
insertion pin locations for each component, as well as other
components within the circuit which it is connected to. This
information is parsed, associated with the corresponding
touch button, and read using a speech synthesizer. The
resulting system allows for input of a Fritzing Diagram, and
output of a 3D model and audio interface. Several challenges
were involved in translating Fritzing diagrams to 3D
representations appropriate for a TUI. Crossed wires are
often present in Fritzing diagrams, but are problematic when
translated to a TUI due to the capacitive nature of our
interaction technique. If two wires are crossed, it may be
difficult for our audio interface to differentiate between the
two wires, resulting in confusing feedback. Our system
addresses this by identifying potential crossed wires and
locating suitable alternatives that maintain circuit
connectivity. Similar issues exist for components that are
positioned close together in the Fritzing diagram. This could
result in a component being difficult to touch in isolation,
again resulting in confusing audio feedback. The
TangibleCircuits parsing algorithm locates components
which are within 2 pins of each other, and considers the
quantum of unoccupied pins surrounding the component. If
such space exists, the algorithm redistributes the components
with 2 pins in between. This ensures that components were
spread-out enough to be recognizable through touch. We
identified that 2 pins were sufficient for our purposes through
a small pilot study with a BVI student. We also determined
the dimensions that smaller components, such as wires and
resistors, should be printed for tactile recognizability. As a
result of this inquiry, we adjusted the wire thickness to 1mm
(scaled larger) and left the component size the same. Scaling
components larger actually resulted in greater confusion for
the participant as well as complicated our parsing
algorithm’s ability to redistribute components effectively.
STUDY 3: FORMAL USER STUDY
In order to understand the effectiveness of TangibleCircuits
for assisting BVI users at understanding sample breadboard
circuits, we conducted a formal user study. The focus of this
evaluation is to understand how TangibleCircuits
complements and contrasts open-source web tutorials at
communicating circuit tutorial implementation. Our study
consisted of two sessions: learning and testing, as well as two
stages: Tangible Circuits and web tutorials. In the learning
session, participants were asked to learn a sample circuit
using either TangibleCircuits or web tutorials modified to
WCAG accessibility standards. The testing session followed
the learning session immediately, in which, participants were
asked to complete two tasks: a component identification task
and an error identification task.
Participants
14 participants (10 female) with varying self-reported visual
impairments (8 self-reported blind) and electronics
educational backgrounds (11 self-reported “none”) were
Figure 5: TangibleCircuits audio interface displayed on a commodity
smartphone overlaid with the 3D printed circuit model.
7
recruited through online advertising, and assistance from a
local organization serving the BVI community. Participants
ranged in age between 27 and 67 with a median age of 47.
Participants were compensated for their time.
Apparatus
Study apparatus included the interactive tactile circuit
diagrams running on an Android smartphone. For the web
tutorial condition, participants were asked to bring their own
laptop equipped with their preferred accessibility tools due
to the common practice of highly customizing BVI screen
reader and screen magnifier interfaces to suit their needs.
Task
Learning Session
For the TangibleCircuits stage, participants were introduced
to our interactive prototype’s functionality and usage. We
briefly explained the audio feedback mechanism and
demonstrated the general use of the device. The audio
feedback for each component contained information
regarding the components placement and connectivity to
other components on the breadboard. We then asked
participants to explore the spatial relationship of
components, and geometry of components using the tangible
and audio feedback of the device. Once participants felt they
had a reasonable understanding of the circuit structure, we
proceeded to the testing session.
For the web-tutorial session, participants were asked to
navigate to an online web tutorial which had modified to
meet WCAG accessibility standards and uploaded to Project
Arduino Hub (Figure 6). To fully bring these tutorials to
accessibility standards, we referenced the Smith-Kettlewell
Technical File (SKTF) for examples of hardware computing
tutorials designed specifically for BVI engineers [4]. The
SKTF is a commonly used reference manual for circuit
descriptions and tutorial descriptions for BVI engineers.
Using this document as a resource, a member of our research
team with a formal background in computer engineering in
collaboration with our experts modified these tutorials to
meet WCAG and SKTF accessibility standards. These
modifications included adding component descriptions,
circuit descriptions, step-by-step written implementation
instructions, and written video caption transcriptions. Each
tutorial contained a list of components needed to implement
the circuit, as well as component descriptions we created
according to [4]. In addition, each tutorial contained a
written step-by-step direction list for assembling the circuit.
Once participants had opened the tutorial, we asked the
participant to read over the tutorial using their screen reader,
screen magnifier, or other accessibility devices. After the
participant felt they had un understanding of the tutorial
content, we proceeded with the testing session.
Testing session
Component Identification: For this task, we presented
participants with a bucket of 17 common electronic
components (e.g. resistors, LEDS, etc.). The bucket
contained only 1 example of each kind of component. We
then asked participants to use their stage apparatus (web
tutorials or TangibleCircuits) as a guide for identifying
components used in the construction of the tutorial circuit.
Participants were asked to read the name of a component in
the component list, and then pull each physical component
out of the bucket one-at-a-time, and state whether or not the
component they held was the target component from the
tutorial. Participants were not told if their identified
component was correct in order to prevent learning-effects
between the two stages of the study. We recorded whether or
not their choice was correct for each component as well as
the time taken to identify the components. After the tutorial
components have been correctly or incorrectly identified, we
proceed immediately to the circuit error identification task.
Circuit Error Identification: We presented the participant
with a completed circuit using physical components on an
unpowered breadboard (Table 1). Each of these physical
circuits were similar to the circuit described in the tutorial
apparatus (web tutorial or TangibleCircuits), but contained 2
errors: a wire error and a component error. A wire error
involves either a misplacement or missing wire, while a
component error comprises a missing or replaced
component. Participants were then asked to use the tutorial
apparatus as a reference for answering three questions
regarding the physical circuit: 1. Is this physical circuit the
Circuit Name Stage/Difficulty Components
Used
Component Error
/ Wire Error
Push-Button
(See Figure
7A)
Web-Tutorial /
Simple
Push-button,
resistor, 4 wires
Removed resistor /
moved power wire 2
pins left
Modified
Mood Cue
(See Figure
7B)
Web-Tutorial /
Complex
Rotary
potentiometer, 2
capacitors, DC
motor, 9 wires
Replaced capacitor
with push-button /
removed wire
Pressure
Sensor (See
Figure 7C)
TangibleCircuit /
Simple
Pressure sensor,
resistor, 5 wires
Removed pressure-
sensor / Moved
ground wire to
opposite side of
resistor
Continuity
Tester (See
Figure 7D)
TangibleCircuit /
Complex
Buzzer, red
LED, 3-pin
switch, resistor,
6 wires
Replace buzzer with
proximity sensor /
removed wire
connecting LED
Table 1: Details of circuits used during error identification task.
Figure 6: Sample Arduino Project Hub tutorial. A) Component list;
B) Descriptions of components inside the component list.
8
same circuit described in the tutorial? 2. If this circuit is
different, how so? 3. How would you modify this physical
circuit to match the circuit described by the tutorial?
interview. It should be noted that the error modality was the
same across stages. In the case of the simple circuits, the two
error modes were removed a component and move a
component. On the complex circuits, the two error modalities
were replacing and removing a component for both the web
tutorial and the TangibleCircuit stage of the study. Once the
participant has answered these three questions for the circuit,
we proceeded immediately to the next phase of testing.
Procedure
Each session was 90 minutes long and documented using
audio and video recording. Participants were assigned to
group A or group B prior to the study. Group A performed
the web tutorial stage first, and group B performed the
TangibleCircuits stage first. This counterbalance was done in
order to eliminate any potential learning-effects that might
result from our study design. Prior to the study, participants
were given a brief introduction to the functionality of a
breadboard and its role as a tool in circuit prototyping. The
session began with a demographic and technology
experience questionnaire. Participants were then asked to
either asked to complete the web-tutorial stage or the
TangibleCircuits stage, depending on their group
assignment. Each participant completed both the learning
session and the testing session for two different circuit
tutorials in both the web-tutorial and the TangibleCircuit
stage. Following [23], each stage contained one simple
tutorial, and one complex tutorial. These 4 circuit tutorials
were the same for all participants, and each participant
examined the same 4 circuit tutorials. Each stage began with
the learning session of the simple tutorial. After participants
had completed the learning session for the simple tutorial, we
proceeded immediately to the testing session, followed by
the learning and testing session of the complex tutorial. Upon
completing the first stage of the study, we introduced
participants to the apparatus (web tutorial or TangibleCircuit
tutorial) to be used in the second stage of the study. We then
immediately proceeded to the learning and testing session for
the second-stage simple circuit, followed by the learning and
testing session for the second-stage complex circuit. After
completing both stages of the study, participants completed
an exit questionnaire and interview.
Data Analysis
For the identification task of both the web tutorial and the
TangibleCircuit stage, success rates of component
identification were recorded, as well as time taken to identify
each of the components. During the circuit error
identification task, error identification success rate, and
correction rate were recorded in addition to time taken to
answer each of the three questions posed during the task. In
addition, audio and video were recorded, transcribed, and
analyzed to evaluate each participant’s understanding of the
circuits composition and functionality. We also collected
qualitative feedback, as well as Likert-scale usability
evaluations as part of the exit interview.
Findings
Overall, participants performed significantly better on the
component identification task and the circuit error
identification task with TangibleCircuits. In addition, our
qualitative findings reflect that participants enjoyed working
with TangibleCircuits more than web-tutorials. In this
section we revisit our design considerations and discuss how
TangibleCircuits services these criteria within our use-case
scenario of classroom and makerspace accessibility tools.
Support Recognizability Results
We concluded from our formative studies that direct 3D
representation of components as well as providing direct
access to component information through touch could better
support recognizability of components than screen-reader
aided web tutorials. On average, participants identified 62%
of the circuit components with the TangibleCircuits
apparatus versus 34% with the web-tutorials. Furthermore, 3
participants who completed the TangibleCircuits stage first
were able to correctly identify the resistor component, but
unable to do so when subsequently completing the web-
tutorial stage. This indicates that overall, geometric
information of the components was better recovered by
participants using TangibleCircuits than web-tutorials.
Furthermore, participants were able to identify 83% of the
wiring and component errors with TangibleCircuits versus
27% with the web-tutorials. This indicates that spatial
information of the circuit was better communicated using
TangibleCircuits as well. Even when using web tutorials as
a guide, participants expressed a preference to walk through
the tutorial using the physical circuit, touching each
component as they progressed. When asked about this,
participants expressed the need for a physical guide to
accompany the online tutorial information. “It's a spatial
thing, even though I am able to tell where the components
are in the tutorial, I would have no idea if they were in the
right spot [on the physical breadboard]” (P1). This indicates
Figure 7: Circuits used during circuit error identification task with
2 errors each. A) Simple web stage circuit; B) Complex web stage
circuit; C) Simple TangibleCircuit stage circuit; D) Complex
TangibleCircuit stage circuit.
9
the importance of tangible communication in understanding
the spatial information of the circuit.
Multi-Modal Feedback Results
Our formative studies also indicated a need for multimodal
forms of communication to mitigate dependency on purely
visual media. “I was surprised how much I was able to
understand just by touching…I was shocked that I actually
could find the errors” (P2). In addition, participants
emphasized that they believed they would be more capable
of completing the circuit tutorial using TangibleCircuits.
However, participants also cautioned that they may not be
able to replicate the circuit using TangibleCircuits due to the
inaccessible nature of the breadboard itself. Participants also
expressed that web-tutorial’s circuit diagrams and circuit
descriptions were not helpful, and that touching the
TangibleCircuit prototyped was more helpful at
understanding the spatial information of the circuit. “ The
diagrams were useless because I could not see them. I would
never be able to complete the steps on my own” (P9). When
asked if the audio feedback or tangible feedback was more
useful for understanding the circuit’s spatial information,
participants insisted that both were equally useful and
necessary. “It was great having audio feedback together with
touch because together they help better identify the pieces. I
am better with touching things” (P11).
Support Understanding of Circuit Structure Results
A key component indicated by our formative studies to
circuit education is the understanding of circuit structural
information such as connectivity. This information is crucial
for identifying circuit errors and debugging, and is often
lacking for BVI students due to reliance on visual media to
communicate this information. We found that identifying the
wiring error and component error were completed with
different degrees of success. As we can see in Figure 8 these
tasks individually were performed more successfully with
TangibleCircuits than the web tutorials, indicating that
structural information of the circuit was also better
communicated using our prototype. In addition, we found
that participants with total blindness performed differently
than those with low vision. We observed that participants
with low vision relied more on the visual diagrams of the
web tutorials to understand the tutorial material, versus the
textual information. These participants had to view the
monitor very closely using a combination of screen
magnifiers, contrast adjustment software, and screen readers
and reported that using the web-tutorials were strenuous on
their eyes. Participants with total blindness used screen
readers exclusively for the web tutorial stage and overall
performed better using the TangibleCircuits device than their
low-vision peers. Furthermore, participants expressed that
they would prefer to use TangibleCircuits over web tutorials
to learn about circuit prototyping. “The audio is real
advantage. I know when I touch something, I’m hearing
information about that thing…I would never be able to do
that with web [tutorials]” (P8). This immediate access to
relevant information based upon touch contributes to
participant understanding of circuit structure by mitigating
the graphic abstraction common to circuit diagrams.
Automated Accessibility Results
Although the automation and design of our tool is intended
to mitigate the labor demanded on instructors, we found that
some material was not encapsulated within the parsed
Fritzing diagrams. This included useful component
descriptions, which we had to manually insert into our audio
interface. We address this issue in further depth in
Limitations and Future Work. Although our results suggest
that TangibleCircuits could be useful for BVI engineers to
understand spatial and geometric information of the circuit,
many users still expressed a need for step-by-step
instructions in order to feel confident they could replicate the
circuit (Figure 9). This reflects that in our current
implementation, not all necessary information could be
extracted through automation, and thus the original tutorial
still served as a useful tool for some users. Thus, we conclude
that TangibleCircuits serves as a supplementary accessibility
tool, but does not completely replace current tutorial
technology. Instead, TangibleCircuits narrows the gap of
accessibility for these users.
Additional Participant Feedback
Participants reacted enthusiastically to the TangibleCircuit
prototype. The results of our 5 point Likert scale (1 meaning
not at all and 5 meaning very much) exit questionnaire
demonstrated that participants found TangibleCircuits to be
easier to use, less frustrating, and less confusing than web-
tutorials (see Figure 9). 5 of the 7 participants with legal
blindness claimed that the circuit diagrams were the most
Figure 8: Average Success Rate of Simple (S) and Complex (C)
Circuit and Wiring Error Identification Task
Figure 9: Averaged Likert between 1 and 5 with 1 meaning ‘not at
all’ and 5 meaning ‘very much’.
10
useful part of the web tutorial. However, we found that these
participants averaged 32% correctness when performing the
identification task and 13% correctness when performing the
circuit error identification task. We noticed that these
participants often strained to use their eyes, and even
commented that this practice was painful and obtrusive. This
indicates to us that these participants were reluctant to trust
a tangible medium because of their default reliance on sight.
Finally, participants gave several suggestions for how
tangible and audio could be better used together for learning
circuit prototyping. We detail these in future work below.
Designing Accessible Hardware Computing Tutorials
Since this work constitutes the first effort to create tangible
systems for BVI within the domain of hardware computing,
we offer the following design insights for further
investigations in this field.
Design 3D Models for Tactile as Well as Visual Use: A
common pitfall during our study was the misidentification of
components which were similar in tactile quality (e.g. wires
and resistors). This is largely due to the fact that 3D modeled
components are designed for visual, not tangible, usability.
Any system that uses 3D modeled parts for communicating
circuit information must carefully consider the tangible
quality of each component and its distinguishability from
other components with similar tactile qualities.
Work With, Not Against, Current Practices: Many
participants expressed insecurity regarding their ability to
replicate a given circuit with step-by-step instructions or
TangibleCircuits in isolation. This is due to the constraint of
having to count pin holes on a traditional breadboard to
check proper component placement. By designing to support
participants’ understanding of circuit structural information,
multimodal feedback fills a knowledge gap within current
circuit prototyping practices, without diminishing the value
of those practices themselves.
Cost Effective Solutions Through Tertiary Users: We found
during our formative studies that a key to the adoption of
accessibility technologies for STEM education lies in their
cost-effectiveness. This is largely due to the lack of resource
access faced by many BVI engineers, as well as educators
potentially not having access to specialized accessibility
tools. By considering tools readily available to potential
tertiary users (3D printers in maker spaces, smartphones,
etc.) we shift the financial burden of creating accessible
education and broaden the inclusivity of classrooms.
Mitigate Information Overload with Gesture: During our
user study, participants suggested that the touch based
interaction could be improved by reconsidering our current
gesture. In our current implementation, the device continues
to read information regarding the selected component to the
participant until the user stops touching that particular
component. 3 participants mentioned that this relayed too
much information, and a multi-tap gesture might work better.
Multi-tap would allow different information to be
communicated about the component each time it is touched.
Furthermore, multi-touch input techniques could also be
helpful for allowing a user to touch 2 components
simultaneously and receive information regarding their
relationship. These considerations of touch input technique
remain a promising avenue for further inquiry into touch and
audio hardware computing tutorial systems.
LIMITATIONS AND FUTURE WORK
The promising results of this initial work indicate many
avenues for future investigation. Although our parsing tool
can automatically construct an audio interface using the
information in the Fritzing diagram, these files often do not
contain all details necessary for BVI users to understand a
given circuit. For this reason, our audio interface required
some manual input of missing information including
component color and usable component names (e.g. “green
wire” vs “wire 5”). However, this problem could be easily
mitigated by embedding this information within the id tags
of the file itself using techniques such as [43]. Furthermore,
TangibleCircuits is suited for small circuits which are not
egregiously complicated. The majority of web tutorial
circuits are simple, suitable for novices to use for learning
fundamentals. It is our vision that more complex circuits
could be explored and implemented using TangibleCircuits
by decomposing large, complicated circuits into smaller,
modular elements which could be integrated to implement
the larger system. Future work will explore algorithmic
techniques to implement this decomposition process.
Finally, in order to ensure the universality of our design, we
intend to deploy a similar user study with sighted users.
CONCLUSION
We present the magnitude of accessibility limitations novice
BVI engineers face in understanding the geometric, spatial
and structural information within the domain of hardware
computing. Through a semi-structured interview with 3 BVI
makers as well as formative studies, we compiled 4 design
considerations to inform the construction of a multimodal
tangible and audio interface for replicating breadboard
circuits called TangibleCircuits. This system comprises an
automatic parsing algorithm which takes a Fritzing Diagram
as input, and renders a 3D model and touch-based audio
interface as output. These two elements are combined to
create our interactive device which fits a capacitive smart-
phone form factor. Our formal user study indicates that
TangibleCircuits mitigates the accessibility gap of web-
tutorials, and is enjoyable for BVI students to use. We
believe BVI users bring valuable perspectives to hardware
computing and push for greater inclusion of their voices and
insights. It is our vision that BVI engineers will design and
construct their own accessibility devices in the future.
ACKNOWLEDGMENTS
We would like to thank Josh Miele, Carol Center for the
Blind, and the Dartmouth Department of Accessibility for all
of their support in this work.
11
REFERENCES
[1] Hackster.io. 2019. Arduino Project Hub. Last
retrieved 18 September 2019 from
https://create.arduino.cc/projecthub
[2] Friends of Fritzing Foundation. 2019. Fritzing Hub.
Last retrieved 18 September 2019 from
https://fritzing.org/projects/
[3] Primo Toys. 2016. Cubetto – Hand on coding for
ages 3 and up. Last retrieved March 14th, 2016 from
https://www.kickstarter.com/projects/primotoys/cube
tt o-hands-on-coding-for-girls-and-boys-aged-3
[4] Smith-Kettlewell Eye Research Institute. 2019.
Smith-Kettlewell Technical File. Last retrieved 18
September 2019 from https://www.ski.org/smith-
kettlewell-technical-file
[5] Web Accessibility Initiative. 2019. Web Content
Accessibility Guidelines. Last retrieved 18
September 2019 from
https://www.w3.org/WAI/fundamentals/accessibility-
principles/
[6] Fraser Anderson, Tovi Grossman, and George
Fitzmaurice. 2017. Trigger-Action-Circuits:
Leveraging Generative Design to Enable Novices to
Design and Build Circuitry. In Proceedings of the
30th Annual ACM Symposium on User Interface
Software and Technology. 331–342.
DOI=https://doi.org/10.1145/3126594.3126637
[7] Jeffrey P. Bigham, Maxwell B. Aller, Jeremy T.
Brudvik, Jessica O. Leung, Lindsay A. Yazzolino,
and Richard E. Ladner. 2008. Inspiring blind high
school students to pursue computer science with
instant messaging chatbots. In Proceedings of the
39th SIGCSE technical symposium on Computer
science education (SIGCSE '08). ACM, New York,
NY, USA, 449-453.
DOI=http://dx.doi.org/10.1145/1352135.1352287
[8] Tracey Booth, Simone Stumpf, Jon Bird, and Sara
Jones. 2016. Crossed Wires: Investigating the
Problems of End-User Developers in a Physical
Computing Task. In Proceedings of the 2016 CHI
Conference on Human Factors in Computing
Systems. 3485–3497.
DOI=https://doi.org/10.1145/2858036.2858533
[9] E. Brulé, G. Bailly, A. M. Brock, F. Valentin, G.
Denis, and C. Jouffrais, “MapSense: Multi-Sensory
Interactive Maps for Children Living with Visual
Impairments,” In Proceedings of the 2016
Conference for Human-Human Factors in Computing
Systems (CHI 2016). ACM, San Jose, California,
USA, pp. 445–457, 07-May-2016.
[10] Z. Cattaneo and T. Vecchi. 2011. Blind vision: the
neuroscience of visual impairment, no. March. MIT
Press. pp.83-89
[11] Daniel Drew, Julie L. Newcomb, William McGrath,
Filip Maksimovic, David Mellis, and Björn
Hartmann. 2016. The Toastboard: Ubiquitous
Instrumentation and Automated Checking of
Breadboarded Circuits. In Proceedings of the 29th
Annual Symposium on User Interface Software and
Technology. 677– 686.
DOI = https://doi.org/10.1145/2984511.2984566
[12] Josh Davis. 2019. IllumiWear: A Fiber-Optic
eTextile for MultiMedia Interactions. In the
Proceedings of the Conference on New Interfaces in
Musical Expression (NIME’19).
[13] Josh Davis, Jun Gong, Yunxin Sun, Parmit Chilana,
Xing-Dong Yang. 2019. CircuitStyle: A System for
Peripherally Reinforcing Best Practices in Hardware
Computing. In Proceedings of the 32nd Annual ACM
Symposium on User Interface Software and
Technology. ACM.
DOI = https://doi.org/10.1145/3332165.3347920
[14] J. Ducasse, A. Brock, and C. Jouffrais, “Accessible
Interactive Maps for Visually Impaired Users,” in
Mobility in Visually Impaired People - Fundamentals
and ICT Assistive Technologies, E. Pissaloux and R.
Velasquez, Eds. Springer, 2017.
[15] J. Ducasse, M. Macé, M. Serrano, and C. Jouffrais,
“Tangible Reels: Construction and Exploration of
Tangible Maps by Visually Impaired Users,” in
Proceedings of the 2016 CHI Conference on Human
Factors in Computing Systems, 2016, pp. 2186–2197.
[16] Caitlin Duncan, Tim Bell, and Steve Tanimoto. 2014.
Should your 8-year-old learn coding?. In Proceedings
of the 9th Workshop in Primary and Secondary
Computing Education (WiPSCE '14). ACM, 60-69.
DOI = http://doi.acm.org/10.1145/2670757.2670774
[17] S. Fleck and M. Hachet, “Making Tangible the
Intangible: Hybridization of the Real and the Virtual
to Enhance Learning of Abstract Phenomena,” Front.
ICT, vol. 3, p. 30, Dec. 2016.
[18] Louis H. Goldish, Harry E Taylor. 1974. The
Optacon: A Valuable Device for Blind Persons. In
New Outlook for the Blind, 68, 2, 49-56, Feb 74.
[19] Michael S. Horn and Robert J. K. Jacob. 2007. Designing tangible programming languages for
classroom use. In Proceedings of the 1st
international conference on Tangible and embedded
interaction (TEI '07). ACM, 159-162.
DOI = http://dx.doi.org/10.1145/1226969.1227003
12
[20] P.A. Ioannou ; C.C. Chien. 1993. Autonomous
intelligent cruise control. Published in IEEE
Transactions on Vehicular Technology. Volume:
42 , Issue: 4 , Nov 1993.
DOI = https://doi.org/10.1109/25.260745
[21] Shaun K. Kane and Jeffrey P. Bigham. 2014.
Tracking @stemxcomet: teaching programming to
blind students via 3D printing, crisis management,
and twitter. In Proceedings of the 45th ACM technical
symposium on Computer science education (SIGCSE
‘14). ACM, New York, NY, USA, 247-252.
DOI = http://dx.doi.org/10.1145/2538862.2538975
[22] Jingyi Li, Son Kim, Joshua A. Miele, Maneesh
Agrawala, Sean Follmer. 2019. Editing Spatial
Layouts through Tactile Templates for People with
Visual Impairments. In Proceedings of the 2019 CHI
Conference on Human Factors in Computing Systems
(CHI '19).
DOI = https://doi.org/10.1145/3290605.3300436
[23] Jo-Yu Lo, Da-Yuan, Tzu-Sheng Kuo, Chen-Kuo,
Jun, Teddy Seyed, Xing-Dong Yang, Bing-Yu
Chen. 2019. AutoFritz: Autocomplete for Prototyping
Virtual Breadboard Circuits. In Proceedings of the
2019 CHI Conference on Human Factors in
Computing Systems (CHI '19).
DOI = https://doi.org/10.1145/3290605.3300633
[24] Stephanie Ludi, Lindsey Ellis, and Scott Jordan.
2014. An accessible robotics programming
environment for visually impaired users. In
Proceedings of the 16th international ACM
SIGACCESS conference on Computers &
accessibility (ASSETS '14). ACM, 237- 238.
DOI = http://dx.doi.org/10.1145/2661334.2661385
[25] M. S. Manshad, E. Pontelli, and S. J. Manshad,
“Trackable Interactive Multimodal Manipulatives:
Towards a Tangible User Environment for the
Blind,” in Proceedings of ICCHP 2012, 2012, vol.
7383, pp. 664– 671.
[26] D. McGookin, E. Robertson, and S. Brewster,
“Clutching at Straws: Using Tangible Interaction to
Provide Non-Visual Access to Graphs,” in
Proceedings of the 28th Conference on Human
Factors in Computing Systems - CHI ’10, 2010, pp.
1715–1724.
[27] Will McGrath, Daniel Drew, Jeremy Warner,
Majeed Kazemitabaar, Mitchell Karchemsky, David
Mellis, and Björn Hartmann. 2017. Bifröst:
Visualizing and Checking Behavior of Embedded
Systems Across Hardware and Software. In
Proceedings of the 30th Annual ACM Symposium on
User Interface Software and Technology. 299– 310.
DOI=https://doi.org/10.1145/3126594.3126658
[28] Orni Meerbaum-Salant, Michal Armoni, and
Mordechai Ben-Ari. 2013. Learning computer
science concepts with Scratch. Computer Science
Education 23(3), 239-264.
DOI=http://dx.doi.org/10.1080/08993408.2013.8320
22
[29] David A. Mellis, Leah Buechley, Mitchel Resnick,
and Björn Hartmann. 2016. Engaging Amateurs in
the Design, Fabrication, and Assembly of Electronic
Devices. In Proceedings of the 2016 ACM
Conference on Designing Interactive Systems. 1270–
1281.
DOI=https://doi.org/10.1145/2901790.2901833
[30] Mitchel Resnick, Fred Martin, Randy Sargent and
Brian Silverman. 1996. Programmable bricks: Toys
to think with. IBM Systems journal, 35 (3.4). 443-
452.
[31] Jaime Sánchez, and Fernando Aguayo. 2005. Blind
learners programming through audio. In Extended
Abstracts on Human Factors in Computing Systems
(CHI EA '05). ACM, 1769-1772.
http://dx.doi.org/10.1145/1056808.1057018
[32] O. Shaer and E. Hornecker, “Tangible User
Interfaces: Past, Present, and Future Directions,”
Found. Trends Human–Computer Interact., vol. 3,
no. 1–2, pp. 1–137, Jan. 2009.
[33] Lei Shi, Zhuohao Zhang Zhejiang, Shiri Azenkot.
2019. Designing Interactive 3D Printed Models with
Teachers of the Visually Impaired. In Proceedings of
the 2019 CHI Conference on Human Factors in
Computing Systems (CHI '19).
DOI=https://doi.org/10.1145/3290605.3300427
[34] Lei Shi Yuhang Zhao Shiri Azenkot . 2017. Markit
and Talkit: A Low-Barrier Toolkit to ' Augment 3D
Printed Models with Audio Annotations. In
Proceedings of the 30th Annual ACM Symposium on
User Interface Software and Technology (UIST ’17).
DOI=https://doi.org/10.1145/3126594.3126650
[35] Lei Shi Yuhang Zhao Shiri Azenkot . 2017.
Designing Interactions for 3D Printed Models with
Blind People. In Proceedings of the 19th
international ACM SIGACCESS conference on
Computers & accessibility (ASSETS '17). ACM.
DOI=https://doi.org/10.1145/3132525.3132549
[36] Lei Shi, Idan Zelzer, Catherine Feng, and Shiri Azenkot. 2016. Tickers and Talker: An Accessible
Labeling Toolkit for 3D Printed Models. In
Proceedings of the 2016 Conference on Human
Factors in Computing Systems (CHI '16). ACM, New
York, NY, USA, 4896-4907.
DOI=https://doi.org/10.1145/2858036.2858507
13
[37] J. Schneider and T. Strothotte, “Constructive
exploration of spatial information by blind users”. In
Proceedings of the fourth international ACM
conference on Assistive technologies - ASSETS ’00,
2000, pp. 188–192.
[38] Ann C. Smith, Joan M. Francioni, and Sam D.
Matzek. 2000. A Java programming tool for students
with visual disabilities. In Proceedings of the fourth
international ACM conference on Assistive
technologies (ASSETS '00). ACM, 142-148.
DOI=http://dx.doi.org/10.1145/354324.354356
[39] Andreas M. Stefik, Christopher Hundhausen, and
Derrick Smith. 2011. On the design of an educational
infrastructure for the blind and visually impaired in
computer science. In Proceedings of the 42nd ACM
technical symposium on Computer science education
i.https://www.codeclub.org.uk/
ii.https://www.codecademy.com/
iii.https://www.quorumlanguage.com/about.php iv.
http://emacspeak.sourceforge.net/
v. http://sonic-pi.net/ (SIGCSE '11). ACM, New
York, NY, USA, 571-576.
DOI=http://dx.doi.org/10.1145/1953163.1953323
[40] Evan Strasnick, Maneesh Agrawala, and Sean
Follmer. 2017. Scanalog: Interactive Design and
Debugging of Analog Circuits with Programmable
Hardware. In Proceedings of the 30th Annual ACM
Symposium on User Interface Software and
Technology. 321–330.
DOI=https://doi.org/10.1145/3126594.3126618
[41] Anja Thieme, Cecily Morrison, Nicolas Villar,
Martin Grayson and Siân Lindley. 2017. Enabling
Collaboration in Learning Computer Programing
Inclusive of Children with Vision Impairments. In
Proceedings of the 2017 ACM Conference on
Designing Interactive Systems (DIS ’17). ACM, New
York, NY, USA
DOI=http://dx.doi.org/10.1145/3064663.3064689
[42] Jeremy Warner, Ben Lafreniere, George Fitzmaurice,
Tovi Grossman. 2018. ElectroTutor: Test-Driven
Physical Computing Tutorials. Proceedings of ACM
Symposium on User Interface Software Technology.
(UIST’18). ACM, New York, NY, USA
[43] Sujeath Pareddy, Anhong Guo, and Jeffrey P.
Bigham. 2019. X-Ray: Screenshot Accessibility via
Embedded Metadata. In The 21st International ACM
SIGACCESS Conference on Computers and
Accessibility (ASSETS ’19). Association for
Computing Machinery, New York, NY, USA, 389–
395.
DOI=https://doi.org/10.1145/3308561.3353808
[44] Xiaoyi Zhang, Tracy Tran, Yuqian Sun, Ian Culhane,
Shobhit Jain, James Fogarty, Jennifer Mankoff. 2018.
Interactiles: 3D Printed Tactile Interfaces to Enhance
Mobile Touchscreen Accessibility. In The 20th
International ACM SIGACCESS Conference on
Computers and Accessibility (ASSETS ’18).
Association for Computing Machinery, New York,
NY, USA
DOI=https://doi.org/10.1145/3234695.3236349